Research map / Image generation / Editing and style transfer
Exploring Unlabeled Faces for Novel Attribute Discovery
Hyojin Bahng, Sunghyo Chung, Seungjoo Yoo, Jaegul Choo
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| discover facial domains from unlabeled images using pretrained features and cluster-conditioned translation | XploreGAN · CelebA attribute transfer | classification accuracy: Blond: 90.2 vs 90.0; Brown: 77.4 vs 86.1; Aged: 90.0 vs 88.4; Male: 99.7 vs 97.5; Female: 99.6 vs 98.0 | Mixed |
| discover domains from unlabeled data rather than train with attribute labels | XploreGAN · CelebA user study | user preference: Hair: 54.7 vs 45.3; Gender: 64.5 vs 14.6; H+G: 89.6 vs 10.4; H+A: 53.1 vs 46.9 | Improved |
| use cluster-summary ASIN instead of instance-conditioned AdaIN (ablation) | XploreGAN · CelebA facial attribute translation | attribute specificity and identity preservation | Improved |
| use BagNet-17 pretrained on ImageNet to extract features (ablation) | attribute clustering · CelebA and FFHQ | discovery of texture attributes | Improved |
| use ResNet-50 pretrained on Stylized ImageNet to extract features (ablation) | attribute clustering · EmotioNet | discovery of shape and expression attributes | Improved |
| increase the number of clusters (ablation) | XploreGAN · unlabeled facial images | attribute granularity | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- StarGAN — StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation (CVPR 2018) · better on 9 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Improving Facial Attribute Prediction Using Semantic Segmentation (CVPR 2017 · similarity 0.86)
- Exploring Explicit Domain Supervision for Latent Space Disentanglement in Unpaired Image-to-Image Translation (TPAMI 2019 · similarity 0.86)
- Scaling-Up Disentanglement for Image Translation (ICCV 2021 · similarity 0.85)
- Consensus-Driven Propagation in Massive Unlabeled Data for Face Recognition (ECCV 2018 · similarity 0.85)
- Generative Unlearning for Any Identity (CVPR 2024 · similarity 0.85)
- FaceA-Net: Facial Attribute-Driven ID Preserving Image Generation Network (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
